1. Field of the Invention
The present invention relates to network devices, circuitry, and architecture. More particularly, the present invention relates to repeater circuits and stacking bus architectures used within 10 MB or 100 MB Ethernet other computer data networks.
2. Description of the Prior Art
Ethernet is an example of a well-known and popular standard for facilitating communication between devices and machines within information networks. Data networks such as Ethernet is so widely used that it is almost impossible to find an office in the United States that does not employ several Ethernet LANS (local area networks) used to facilitate its day-to-day business operations.
Typically, within a data network, devices are connected to one another via a wire, such as category 5 (CAT 5) or twisted pair 10 BASE-T wires that are flexible and allow for easy cable pulling through building walls, ceilings, etc. Ethernet relies on a communication protocol called Carrier Sense Multiple Access Collision Detect (CSMA/CD). Each station within an Ethernet network is connected to a single wire used to both transmit and receive data. The Carrier Sense of CSMA/CD means that before transmitting data, a station must check the wire to see if any other station is already sending data. Accordingly, a station will typically only send data when the LAN appears to be idle (i.e., no signals coming in).
Cables used within Ethernet networks have physical limitations that must be accounted for in a network architecture. For example, an Ethernet station in a 10 MB Ethernet network sends data at a rate of 10 MB per second. At this rate, a bit travels approximately 100 feet down a network cable before the second bit is sent. So, if two stations are located 250 feet apart, for example, and both begin transmitting at the same time, then each station will be in the middle of sending its third bit before the signal from each other reaches the other station. When two signals are sent onto the same network segment at the same time, a collision occurs and the signals are lost. Therefore, there is a need for Collision Detection.
Another problem caused by the physical nature of cable causes within the network topology is signal attenuation. The resistivity of copper cable or wire causes signals to attenuate over a certain distance such that an Ethernet station receiving a data packet may not be able to accurately read each and every bit of the data pack (an Ethernet packet has a well-known structure that includes a preamble, which network devices capture and use to determine what to do with a packet; signal attenuation can make it difficult or impossible to read the preamble). Accordingly, a well-known device called a repeater is used within a network to restore the signal and remove the effects of amplitude distortion caused by signal attenuation and timing distortion caused by jitter, which the signal experiences as it propagates through each network segment.
In its most basic form, a repeater receives data on a physical port and repeats to all of its other ports except the active receiver port on the repeater, restoring signal amplitude and timing on the retransmitted data packets. As explained above, a collision occurs when signals are sent by multiple machines on the same wire. Therefore, another common function that a repeater performs is Collision Detection. If the repeater detects receive activity from two or more ports, this constitutes a collision (i.e., two machines are attempting to send a signal at the same time), and the repeater will send a jam pattern on all ports, including the active receive ports. Reception and retransmission of signals and packets are closely specified in the section, “Repeater Units for 10 MB per Second Base Band Networks” of the IEEE 802.3 standard.
While repeaters are required for Ethernet networks, they introduce some other effects that must be accounted for when building large networks. One such effect is delay. Repeaters introduce delay into the network signal as it propagates signals from one port to another. This delay must be factored into the overall roundtrip delay of the network. Another effect is referred to as “interpacket gap shrinkage” (IPG shrinkage). The main cause of IPG shrinkage is the variability of the delay path through the repeater for back-to-back packets.
In order to understand the effective IPG shrinkage, consider the example in which two packets are issued from a transmitting station with minimum IPG. When the first packet reaches a repeater, the repeater will take a certain amount of time to recognize the signal and pass it to its other ports, therefore introducing a delay. As the delay between packets varies even slightly, as packets are transmitted from repeater to repeater, the gap between packets can be shortened. If the IPG becomes too small, repeaters may not be able to reacquire lock to the incoming packet (by reading the preamble) and may therefore decode some of the packet data incorrectly. The basic outcome of both the repeater delay and the IPG shrinkage issues is that the number of repeaters permitted in the end-to-end path of the network must be restricted.
Accordingly, there is need for expanding the collision domain of repeaters and for minimizing IPG gap shrinkage and delay issues caused therefrom. One way of expanding collision domain is to create a repeater with more ports. However, in order to increase the number of physical ports, the number of repeater circuits (e.g., repeater chips) in a repeater must be increased. A second way of increasing the size of a collision domain is to link repeaters together in such a way that all the ports of each repeater are in the collision domain. One way of linking multiple repeaters together is via circuitry commonly referred to as a backplane. Backplanes allow repeater manufacturers to cascade multiple repeater circuits into a single hub. In order to allow multiple repeater circuits in a system to behave as a single hub, the repeaters must pass collision information in addition to rounding data and clock signals.
Thus, there is a need for new and improved systems and methods for integrating multiple repeaters (and repeater chips) into a single collision domain. Such systems and methods should be able to handle high-speed, low-speed, or mixed-speed management interconnections between repeaters. Also, such systems and methods should be highly versatile with low cost and ease of design.